Impact of Simulated Climate Data on Wind Power Prediction and Long-Term Grid Planning

This paper assesses the impact of incorporating climate data into long-term power system planning, using simulated wind speed data from Texas as a case study. Two experiments are conducted. First, we evaluate the quality of wind speed time series data obtained from climate model simulations and how it varies with spatial resolution. Our analysis suggests that both high- and low-resolution simulated climate data generally align with the probability distribution of historical data, but high-resolution climate data is able to capture extreme events more accurately. Second, we employ simulated climate data for time-series prediction of daily, weekly, and monthly wind power production. The findings caution against the hasty adoption of climate data for time-dependent prediction, as observations indicate minimal impact on shorter prediction intervals like daily and weekly averaged power generation. The results suggest that the integration of climate data may not provide substantial improvements in forecast accuracy for shorter intervals, under-scoring the need for careful consideration and further research when incorporating climate data into forecasting models. Code availability github.com/fatemehdoudi/Climate4Grid